rch

You can offload builds by running the rchd daemon, defining remote hosts in a workers.toml configuration file, and installing the Claude hook to intercept and route compilation commands to those remote workers.

It uses smart worker selection based on availability, priority, and project locality to maximize throughput. This ensures multiple AI coding agents or CI jobs can trigger concurrent builds that are parallelized across the worker fleet.

No, Docker is not required. The system specifically supports Dockerless SSH-based execution, allowing you to route and execute compilation commands directly on remote workers over SSH without containerization.

The system uses a fail-open design, meaning if remote execution fails, it automatically falls back to local builds. This ensures developers and AI agents are never blocked by remote worker unavailability or execution errors.

The system uses project affinity and incremental builds to reuse cached assets. By selecting workers based on project locality, it ensures that previously compiled assets on the remote worker are reused, avoiding redundant file transfers.

You can validate your setup by running the `rch doctor` command. This checks that the rchd daemon is running, your workers.toml configuration is valid, and the Claude hook integration is correctly installed.